https://juttu-s.github.io/saikiran_juttu.github.io/tags/perception/PerceptionHugo Blox Builder (https://hugoblox.com)en-usThu, 01 May 2025 00:00:00 +0000https://juttu-s.github.io/saikiran_juttu.github.io/media/icon_hu7729264130191091259.pnghttps://juttu-s.github.io/saikiran_juttu.github.io/tags/perception/https://juttu-s.github.io/saikiran_juttu.github.io/details/robotics-coop/Thu, 01 May 2025 00:00:00 +0000https://juttu-s.github.io/saikiran_juttu.github.io/details/robotics-coop/<h2 id="overview">Overview</h2> <p>Worked at Northeastern University on:</p> <hr> <h3 id="sensor-integration">Sensor Integration</h3> <p>Integrated <strong>3D LiDAR and Intel RealSense</strong> sensors with the <strong>Scout Mini Rover</strong> via CAN protocol in ROS 2, deployed on <strong>Jetson AGX Orin</strong> for synchronized perception, sensor fusion, and reliable HW–SW communication.</p> <p><strong>Physical Setup</strong> <figure > <div class="flex justify-center "> <div class="w-100" ><img src="https://juttu-s.github.io/saikiran_juttu.github.io/saikiran_juttu.github.io/uploads/Robot.jpeg" alt="Scout Mini Rover Setup" loading="lazy" data-zoomable /></div> </div></figure> </p> <h3 id="slam-implementation">SLAM Implementation</h3> <p>Implemented ROS 2-based <strong>SLAM systems</strong> using:</p> <ul> <li>RTAB-Map</li> <li>LIO-SAM</li> <li>Visual-Inertial Odometry (VIO)</li> <li>SLAM Toolbox<br> In C++/Python with Nav2 for improved localization in GPS-denied environments.</li> </ul> <p><strong>RTAB-Map Demo:</strong></p> <video controls width="100%" muted autoplay loop> <source src="https://juttu-s.github.io/saikiran_juttu.github.io/saikiran_juttu.github.io/uploads/RTAB_demo.mp4" type="video/mp4"> </video> <p><strong>LIO-SAM Screenshot:</strong></p> <p> <figure > <div class="flex justify-center "> <div class="w-100" ><img src="https://juttu-s.github.io/saikiran_juttu.github.io/saikiran_juttu.github.io/uploads/lio_sam.png" alt="LIO-SAM Visualization" loading="lazy" data-zoomable /></div> </div></figure> </p> <hr> <h3 id="uav-motion-planning">UAV Motion Planning</h3> <p>Developed <strong>waypoint navigation and motion planning</strong> pipelines for <strong>Crazyflie and DJI Tello</strong> using the <strong>OptiTrack motion capture system</strong>.</p> <ul> <li>Automated takeoff</li> <li>Precision landing</li> <li>Closed-loop control with state estimation</li> </ul> <hr> <h3 id="semantic-perception-with-vlms">Semantic Perception with VLMs</h3> <p>Exploring integration of <strong>Vision-Language Models (VLMs)</strong> to:</p> <ul> <li>Enhance indoor semantic navigation</li> <li>Reduce localization drift</li> <li>Enable autonomous scene understanding</li> </ul>